DataStax has moved to repair bridges with the Apache Cassandra community with the unveiling of a Kubernetes operator and a commitment to dramatically increase its contributions to the open source NoSQL database.
The move comes almost six months after the company waved goodbye to former CEO Billy Bosworth and brought in former Google exec and ex- Apigee CEO Chet Kapoor as his replacement. Kapoor has subsequently recruited further execs steeped in Google’s Kubernetes and cloud approach.
The Kubernetes project describes operators as software extensions to Kubernetes that make use of custom resources to manage applications and their components. Datastax said the operator delivered benefits including zero lock-in and global scale. The launch coincided with the general availability of DataStax Enterprise 6.8, which includes native Kubernetes support, and Graph in Cassandra.
Given DataStax has built its business around Cassandra, it might seem peculiar that it hadn’t already delivered a Kubernetes operator. Well, actually it had, just not for Apache Cassandra, which reflected its previously distant relationship with the project, in favour of its own DataStax Enterprise platform.
Since October’s upheaval, said DataStax’ VP of developer relations Patrick McFadin, “We’ve gotten into this mode where we’re reaffirming our mission to make Apache Cassandra the default database for people building applications.”
“But more importantly, we’re watching the world and infrastructure and… we see that Apache Cassandra really should be the default database when you’re working with Kubernetes. So a part of that commitment is we have a Kubernetes operator and a Management API,” he said.
Regarding its previous chilly attitude to Apache Cassandra, McFadin said, “We made some bad choices. And we’re making amends. But you can’t just say we’re going to change things. We have to do things.”
So, the firm has assigned a whole engineering team to Apache Cassandra, he said, which it is looking to recruit further members for.
It was also engaging more closely with other teams that had developed operators. “We’re not just going to release this and say, we think this is the one true operator,” he said.
The bigger strategy is to establish Cassandra as the default data platform on Kubernetes. As the industry shifted through DevOps and cloud native, Kubernetes was now established as the application tier, McFadin said.
“However, we don’t have a cloud native data solution yet…It’s a different class of problem.”
Over the next decade, he predicted that Cassandra could become the default cloud native database. “Why Cassandra? Because it fits into the world of cloud native. It’s that elastic scale, the way that it is distributed and resilient. It can take a lot of abuse.”
The vendor began offering support for Apache Cassandra back in December through its Luna service. Earlier this month it took over Apache Cassandra consulting firm The Last Pickle, which had previously released open source tools such as Cassandra Medusa and Cassandra Reaper.